Clarification if it's allowed to call `fmi3Set` in co-sim on continuous states in event mode

[elmir-nahodovic]

Hi,

my initial thought is no but I would like to have it clarified.

In fmi 3 spec, State: Event Mode (https://fmi-standard.org/docs/3.0.2/#EventMode) it says that fmi3Set is an allowed function call for exposed continuous-time states with reinit = false. Now, reinit is an attribute only used to model exchange. So I assume that for co-simulation you should disregard the sentence "and continuous-time states with reinit = false".

Is that correct?

[elmir-nahodovic]

I now see that this has been discussed in several issues so maybe it can be closed:

#1290 #1950

Moreover the concept of setting states in co-sim between doSteps has also been discussed in #94 (although for fmi 2), there are even articles from Modelica conferences:

2014: "Nonlinear State Estimation with an Extended FMI 2.0 Co-Simulation Interface"
2020: "Nonlinear State Estimation with FMI: Tutorial and Applications"

[chrbertsch]

FMI Design Meeting:
Elmir: I will propose a clarification. And create a ticket for allowing this (giving a rationale, could be in a layered standard or minor release)
Pierre: Same reinit for ME and CS?
Klaus: do we allow adding attributes?
Pierre: we have not said anything about that. Could be possible in a different namespace. Then a schema check will not fail.